Digital stethoscope for detecting a respiratory abnormality and architectures thereof
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11696703-B2
Publication Date
2023-07-11
Expiration Date
Abstract
Embodiments disclosed herein improve digital stethoscopes and their application and operation. A first method detects of a respiratory abnormality using a convolution. A second method counts coughs for a patient. A third method predicts a respiratory event based on a detected trend. A fourth method forecasts characteristics of a future respiratory event. In a fifth embodiment, a base station is provided for a digital stethoscope.
Core Innovation
The invention describes a digital stethoscope for detection of a respiratory abnormality. A microphone detects an auditory signal, and an FPGA generates an auditory spectrogram based on the auditory signal. The FPGA performs a spatial convolution procedure on the auditory spectrogram to generate one or more convolution values representing a compressed version of a matrix portion of the auditory spectrogram.
The spatial convolution procedure is trained to generate one or more features for detecting the respiratory abnormality, and one or more trained weights are applied to the one or more convolution values. The FPGA generates a classification value by applying the one or more weights to the one or more convolution values, and the classification value indicates whether the auditory signal includes the respiratory abnormality.
The classification value is transmitted to a display unit for output. The document also describes cochlear filtering using one or more cochlear filters with a cochlear model, filtering of the auditory signal using a frequency-domain retention/subtraction relationship with time-varying integers, and convolution-related computation paths including storing intermediary convolutional matrix values for successive calculations.
Claims Coverage
The independent claims identified in the provided material are centered on a single core independent claim. Five inventive features are extracted from this claim as the main coverage basis.
Auditory signal to auditory spectrogram on an FPGA
The FPGA receives, from the microphone, the auditory signal and generates an auditory spectrogram based on the auditory signal.
Spatial convolution compression of auditory spectrogram portions
The FPGA performs a spatial convolution procedure on the auditory spectrogram to generate one or more convolution values, wherein the one or more convolution values represent a compressed version of a matrix portion of the auditory spectrogram.
Trained spatial convolution features and trained weights
The spatial convolution procedure is trained to generate one or more features for detecting the respiratory abnormality, and the FPGA applies one or more weights to the one or more convolution values, wherein the one or more weights are trained for detection of the respiratory abnormality.
Classification value indicates respiratory abnormality
The FPGA generates a classification value based on applying the one or more weights to the one or more convolution values, wherein the classification value indicates the auditory signal includes the respiratory abnormality.
Transmit classification value to a display unit
The classification value is transmitted to a display unit for outputting the classification value, and the display unit displays the classification value.
The inventive coverage is directed to an FPGA-based pipeline that converts an auditory signal into an auditory spectrogram, performs spatial convolution that compresses a spectrogram matrix portion into convolution values, uses trained convolution features and trained weights to generate a classification value indicating respiratory abnormality, and outputs that classification value on a display unit.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?